Effects of Oscillatory Flow on the Nucleation and Crystallization of

Sep 21, 2011 - Synopsis. In this study, we investigate the effects of flow, specifically intermittent oscillatory flow, on protein (insulin) crystalli...
8 downloads 13 Views 3MB Size
ARTICLE pubs.acs.org/crystal

Effects of Oscillatory Flow on the Nucleation and Crystallization of Insulin Published as part of a virtual special issue of selected papers presented at the 9th International Workshop on the Crystal Growth of Organic Materials (CGOM9). Jose V. Parambil, Marc Schaepertoens, Daryl R. Williams, and Jerry Y. Y. Heng* Surfaces and Particle Engineering Laboratory, Department of Chemical Engineering, Imperial College London, London SW7 2AZ, United Kingdom ABSTRACT: Protein crystallization has immense potential to be used for separation and formulation applications on an industrial scale. Crystallization under static conditions is often limited by the diffusion of molecules from the bulk solution to the growing crystal surface. This results in a lower overall yield of the product and longer crystallization periods. Hence, protein crystallization under flow conditions has attracted attention with the capability to improve the convective protein transport, reducing the crystallization time and improving yield. In this study, we investigate the effects of flow, specifically intermittent oscillatory flow on protein (insulin) crystallization. It was observed that the nucleation and crystal yield is very much influenced by the flow conditions and herein we present some key observations in insulin crystallization conducted under intermittent oscillatory flow. For oscillatory flow velocities ranging from 6 to 16 mm/min at a frequency of 1 cycle/min, a considerable increase in nucleation has been observed with an increase in flow rate resulting in the formation of large number of crystals in tubes with flow compared to the stationary tubes. Also 50% yield for flow crystallization was achieved in contrast to only 24% yield for the stationary growth over the 48 h period. The flow condition was thus found to affect the number and size of crystals. In addition, the intermittent flow pattern utilized in this study helps to separately understand the influence of flow on nucleation and crystal growth. The flow strategy proposed herein could potentially be utilized to optimize crystallization processes for proteins so as to be used in downstream separation and formulation of products.

’ INTRODUCTION Crystallization has been utilized as a major purification process for small organic molecules for many years now. The pharmaceutical industry uses crystallization as a means of purification and formulation.1,2 Similarly, purification of specific proteins and demonstration of their purity were motives for the development of protein crystallization in the latter half of the 19th century. Further modifications and development shifted the focus of protein crystallization as a tool in understanding the properties and the catalytic activity of enzymes. Since the late 1930s, protein crystallization evolved itself with the primary objective of supplying better crystals for the X-ray diffraction studies.3 Even at present, this remains one of the most important aspects of protein crystallization studies around the globe. With the advent of new scientific realms, the application of protein crystallization is slowly expanding toward biopharmaceutical process designing, formulation of biologics,4 and development of nanoporous media for size exclusion chromatography.5 However, no comprehensive theory or fundamental database is yet developed that can explain and understand the complexities of protein crystallization.3 Most of the difficulties and gray areas in protein crystallization that were enlisted by Rosenberger6 in 1986 still continue to persist in this r 2011 American Chemical Society

field. Typically, optimization of the crystallization conditions is still derived mostly from experimental evidence rather than theoretical approaches.7,8 In search of answering the complex questions in protein crystallization, researchers try to explore crystallization in conjunction with various other facilities and research such as the use of International Space Station9 and genomic technologies to create proteins that are easier to crystallize.10 Most of the developments in protein crystallization had focused on producing high quality crystals from small volumes of protein solution under static conditions.11 Protein crystal growth under quiescent conditions has been reported to be limited by solute diffusion from the bulk of the fluid to the actively growing crystal surface, where the molecules are incorporated into the crystal lattice.12 This has been demonstrated experimentally and supported by model predictions.13 Hence, improving the transfer of protein molecules from the bulk to the crystal surface should in theory result in an increase in the rate of crystallization due to shifting from a diffusion limited to kinetic limited growth regime. Received: January 31, 2011 Revised: August 12, 2011 Published: September 21, 2011 4353

dx.doi.org/10.1021/cg200158z | Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design A thorough understanding of the underlying mechanism and effects of flow modes on protein crystallization would be of great use resulting in better yields and control over the process. With this in mind, flow patterns produced using rotary shakers14 and alternating electric and magnetic fields15,16 have been investigated for improving crystallization in the microscale. Liquid pumps or thermal gradients17,18 have also been utilized to produce the required flow of crystallizing solution. In the crystallization of chicken egg-white lysozyme (CEWL), Roberts et al. have reported that by using a peristaltic pump configuration to obtain a continuous flow of the protein solution through a crystallizing capillary, larger crystals, and narrower crystal size distribution can be obtained in comparison to stationary crystallization.19 Roberts et al. emphasized that protein flow with and without recirculation can enhance crystal growth and is dependent on the flow velocity of the solution through the capillary. Roberts et al. also noted that the continuous flow in their study has helped only in improving the crystal growth rate and did not appear to affect the nucleation process. Though the study by Roberts et al. has opened a promising realm in protein crystallization with increased crystal growth and narrow distribution of crystal size, not every protein would behave like CEWL in a continuous flow pattern. When insulin is used in the flow arrangement as explained by Roberts et al., the protein denatures within a time span of about half an hour and thereby adversely affected the process. This was not completely unexpected as most proteins are biomolecules that are highly sensitive to shear and other physical strains such as temperature and pressure changes. Also, when working with a protein solution that might crystallize under room temperature, a circulatory flow system might cause the crystallization of the solute protein throughout the flow channels making it difficult to control the system and produce good quality crystals. Hence, a flow system that can be contained within a confined volume and at the same time being capable of providing the required convective motion is required for encountering these hurdles. Oscillatory motion of liquid within the capillary tubes is a promising solution in this context. Oscillatory motion of liquid solution for separating the aqueous components based on enhanced diffusion has been studied since late 1960s.20 Effects of oscillatory flow on crystallization of organic molecules21 23 and protein separation using ultrafiltration24 have also been investigated, but oscillatory flow mode in capillary has never been reported to be used in protein crystallization. The flow field produced by oscillatory fluid motion is unique compared to the conditions in microstirring methods with sittingand hanging-drop vapor diffusion techniques used by Adachi et al.,25 or to the continuous flow used by Roberts et al.,19 or solution motion induced using electric, magnetic, or mechanical methods by other researchers15,16,26 to study protein crystallization. The velocity profile produced by oscillatory flow cannot be achieved in most other simple fluid flow systems. It is found from this study that this uniqueness in the fluid dynamics of oscillatory flow mode can have a profound influence on protein crystallization that has not been observed in any other methods. Hence, the effect of oscillatory flow mode on protein crystallization may provide unique opportunities to understand the process and open new perspectives on approaching the bigger problems. In addition to the oscillatory motion of the protein solution, an intermittent flow pattern is also reported in this study. In an intermittent flow study, the crystallizing solution is subjected to various durations of flow and static conditions at regular intervals in a specific order. This intermittency allows segregation of the

ARTICLE

Figure 1. Crystallization cell and piston pump attached to one of the capillaries for oscillatory flow.

effect of the oscillatory flow motion on the nucleation and growth of the protein crystals. As authors’ understand, it is for the first time that such an approach have been utilized in studying the effect of flow on protein crystallization. This study investigates flow effects on protein crystallization, focusing on the effect of oscillatory flow on nucleation and crystal growth. This study is a pioneering step in utilizing intermittent oscillatory flow in capillaries for understanding and controlling the process of protein crystallization.

’ EXPERIMENTAL METHODS Preparation of Insulin Solution. Citrate buffer (0.48 M) at pH 6 was prepared in deionized water using anionic trisodium citrate (Merck) and cationic disodium citrate (BDH Chemicals). The precipitation buffer/precipitant solution was prepared by dissolving zinc sulfate (Sigma-Aldrich) to a concentration of 1% (w/v) in a 10% (v/v) acetone (BDH Chemicals) in 0.48 M pH 6 citrate buffer solution. Stock insulin solution (Sigma-Aldrich, I9278, lot no. 119K8407) containing recombinant human insulin at 5 mg/mL was used in this study. The protein solution was mixed with an equal volume of precipitant solution at room temperature (20 22 °C) to give the crystallizing solution with an initial concentration of 2.5 mg/mL. Flow System. A programmable eight-channel syringe pump, Aladdin8000 (World Precision Instruments Inc., USA) was used to produce the oscillatory flow in the capillary tubes. Glass syringes (2, 10, and 20 mL, Sanitex, Switzerland) were used to obtain the required flow velocities. The syringes were connected to the capillary tubes using tygon tubing (SC0021, Ismatech) of 2.06 mm internal diameter. The entire system was set up as shown in Figure 1. One oscillation cycle of the syringe pump consisted of a pumping phase and a withdrawing phase. This corresponded to one forward and backward movement of insulin solution inside the capillary tubes. The pump settings and the syringes were adjusted suitably to obtain oscillatory flow of various velocities with a frequency of 1 cycle per minute. The flow velocity of insulin solution was measured by monitoring the movement of the liquid meniscus within the capillary tubes. Insulin Crystallization. Crystallization of insulin was carried out in capillary tubes at 10 °C. The soda glass capillaries of 2 mm internal diameter (Bilbate Ltd., UK) were maintained at the desired temperature using a GR 150 digital high performance stirred thermostatic water bath (Grant Instruments Ltd., UK) connected to the crystallization cell as shown in Figure 1. The insulin solution prepared at room temperature was fed into the capillary tubes after the crystallization cell was equilibrated at 10 °C. This enabled us to maintain the same temperature from beginning until end of the experiment. The crystallization process was monitored using a reflective microscope (Olympus, BX51M) with a 4354

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design

ARTICLE

Figure 2. Nucleation enhancement of insulin crystals with flow rate. Images after 48 h in (i) stationary tube and in capillaries with flow at (ii) 6 mm/min, (iii) 8 mm/min, (iv) 12 mm/min, and (v) 16 mm/min. digital camera (Olympus, DP70) connected to a computer with an image capture software Analysis 5.0 (Soft Imaging Systems). In the initial set of experiments, only the oscillatory flow velocity of the solution was varied and other factors including temperature, initial protein concentration, and the oscillating frequency were maintained constants. Control tubes were maintained at static condition while flow was provided in others. In the latter set of experiments with intermittent flow pattern, the duration of flow and static conditions in the experimental tubes were varied while keeping other factors a constant. In this case, two set of control tubes, one under static condition and other with oscillatory flow throughout the experimental duration were maintained. Crystal Yield. The crystal yield was measured using modified Bradford’s method similar to as described by Bergeron et al. 27 using a

UV visible absorption spectrophotometer (Lambda35, Perkin Elmer). The supernatant solution from various capillaries are collected into glass vials after 48 h of crystallization and mixed with appropriate volumes of Bradford’s reagent. Twenty minutes of incubation time is allowed prior to the measurement of absorbance at 595 nm. Components in the solution other than the protein were found not to interfere with the measurement of protein concentration using this method. The concentration of protein in the supernatant solution is then found out against the absorbance in standard plot initially prepared with known concentration of protein solutions using the same procedure. Since the protein concentration in the supernatant solution represents the amount of insulin that was not crystallized during the run, this concentration value was then used to calculate the amount of crystals formed and the corresponding crystal yield. 4355

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design

ARTICLE

Figure 3. Insulin crystals grown for 36 h in (i) stationary control, DS1, and tubes started under stationary conditions and flow initiated after (ii) 4 h, D1, (iii) 8 h, D3, (iv) 12 h, D5, and (v) flow control, DF1. Flow rate in DF1, D1, D3, and D5 were 12 mm/min at a frequency of 1 cycle per minute.

’ RESULTS The observations of insulin crystallization with oscillatory flow could be approached in two different aspects regarding the influence of flow on nucleation and on overall crystal yield. Both these factors are considered in the following discussion. Effect on Nucleation. Insulin produced a higher number of crystals with flow in comparison to the stationary tubes. The oscillatory flow was found to enhance nucleation in all flow velocities under consideration. Experiments conducted at various flow velocities also found that a strong correlation exists between crystal nucleation, growth rate, and flow rate. Upon increasing the flow rate from 6 mm/min to 16 mm/min, at a constant

frequency of 1 cycle per minute, it was observed that the number of crystals and crystal yield increased with increasing flow. Figure 2 gives a representative view of the growth of insulin crystals under various flow rates after 48 h. As we move from Figure 2i v in increasing order of oscillating flow velocity, we observe that there is a corresponding increase in the number of crystals. This is an indication that insulin nucleation is enhanced by the oscillatory flow rate at the frequency of 1 cycle per minute within the velocity range of this study, that is, 6 16 mm/min. In order to explicitly prove this effect of oscillatory flow on nucleation, delayed start of flow was utilized. In delayed-start experiments, flow was initiated in different capillary tubes at 4356

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design

ARTICLE

Figure 4. Variation of maximum shear rate vs average flow velocity.

different times after crystallization had commenced. That is, crystallization was started in these tubes under stationary conditions, and flow was then provided after a specific time delay of 4 h (tubes D1 and D2), 8 h (tubes D3 and D4), and 12 h (tubes D5 and D6). Hence, the crystallization in these tubes started under stationary conditions but was then shifted to oscillatory flow conditions with a fixed flow velocity. Two sets of control tubes were maintained with two capillaries each one set at stationary condition (DS1 and DS2) and another set under permanent oscillation (DF1 and DF2) over the whole course of each experiment. This allowed for segregation of crystallization behavior with respect to nucleation and growth pattern of insulin crystals within the scope of the two conditional extremes utilized in this study — no-flow and oscillatory flow throughout. In the control tubes under stationary conditions (DS1 and DS2), the crystals that grew in the 48 h time period were few in number but with a size range of 300 450 μm. In the control tubes with flow throughout (DF1 and DF2), a large number of crystals was observed and continued to grow for the 48 h window. The crystal size distribution that was observed in the delayed-start tubes (D1 D6) was interesting in context of this study. As expected, the crystal growth in the delayed-start tubes remained similar to that within the stationary control tubes until flow was introduced. Once, the flow was provided, a larger number of small crystals were observed growing along with the previously observed crystals. That is, more crystals nucleated after flow was given. This is observed in Figure 3. In comparison with DF1 and DF2, the size distribution of crystals in D1 D6 stands in sharp contrast with a few very large crystals surrounded by a lot of smaller crystals. This is clearly observable in panel (iv) of Figure 3 as there are a large number of crystals of size less than 100 μm alongside a few crystals that are above 200 μm in size. This large size difference is strong evidence that the smaller crystals nucleated after flow had been provided, while the larger crystals had nucleated and grown during the stationary phase and continued to grow under flow conditions. Thus, we propose that oscillatory flow has a promoting effect on nucleation of insulin crystals. Shear Rate and Nucleation. Solution shear flow in protein solution droplets has been reported to increase crystal nucleation.28 Also, oscillatory shear flow is reported to produce ordered arrangement in colloidal structures with spheres of less than a micrometer in diameter.29 Even micelles as small as 6 nm have been found to form hexagonal symmetry in macroscopic dimensions

Figure 5. Shear rate distribution profile in fluid with average flow velocity of 16 mm/min at the time when the fluid experiences the maximum shear rate.

under the influence of oscillatory shear flow.30 The hydrodynamic radius of insulin hexamer has been reported in the range of 5 6 nm.31 Also, the high density clusters that are suggested to be the precursor of protein crystal nuclei in a two-step nucleation process are reported to be on the order of hundreds of nanometers.32 Hence, we suggest that the improvement in the nucleation of insulin crystals could be caused by the shear effect on solution inside the capillary tubes as suggested by Anita et al.28 and the ordering of the insulin hexamers or the dense liquid clusters occur under the effect of oscillatory flow. Thus, the combined effect of the shear and oscillatory fluid motion results in the formation of a large number of crystal nuclei once the flow is initiated. But an extensive quantification of the nucleation and shear rate in oscillatory flow would be required to find out the optimal shear required for the nucleation of insulin crystals under this condition. In order to quantify the shear rate developed in this study, fluid flow modeling using COMSOL (Multiphysics Modelling and Simulation Software) was utilized. Using the laminar flow condition in a capillary tube as in the experiment and providing the suitable initial and boundary conditions, the velocity distribution and the shear rate developed within the oscillatory flow was computed. This maximum shear rate for various flow rates used in the study are plotted in Figure 4. Also, the shear rate distribution profile within the fluid at the time when this maximum shear rate is experienced in 16 mm/min flow velocity is provided in Figure 5. The results showed that the shear rate in the fluid varied from 0.01 s 1 to 0.62 s 1. The maximum shear rate increased from 0.23 s 1 for flow velocity of 6 mm/min to 0.62 s 1 for 16 mm/min velocity. This shear rate is within the range where ordered structural arrangements of colloidal particles have been observed.33 This provides further reliability in relating nucleation enhancement of insulin by shear flow to the ordered arrangement of colloids under similar conditions. That is, under the influence of the computed shear rate, the insulin hexamers or high density liquid clusters of insulin forms ordered arrangements in shear planes similar to colloidal arrangement. Once this small-scale ordered arrangement is established, the nucleation event could follow rapidly, resulting in a large number of nuclei throughout the capillary. 4357

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design

Figure 6. Percentage yield of insulin crystals against flow rate.

Apart from shear rate, studies on colloidal structures show that the frequency of oscillation could have an impact on the crystallization behavior of proteins. Thus, controlling the flow velocity and frequency in an oscillatory flow protein crystallization would provide us with a valuable tool in controlling the event of nucleation and thereby resulting in controlled crystallization as such. Effect on Crystal Yield. Apart from increasing the number of crystals or obtaining a narrow size distribution of crystals, it is also important to observe efficiency of flow process in recovering protein from solution. It would be useful if we can obtain a comparative analysis on the amount of protein that can be recovered from the crystallizing solution between a stationary experiment and flow condition. This will be helpful in determining if the new method would save time and energy in further works. Hence the effect of oscillatory flow on overall crystal yield was investigated and compared with that from stationary crystallization. As with nucleation, crystal yield studies also exhibit a direct correlation with the flow rate of the crystallization solution. The yields for various flow rates are plotted in Figure 6. Under stationary conditions for a time period of 48 h, the insulin crystal yield for our conditions was observed to be 23.5 ( 2.9%. In the case of tubes with oscillations, for the same duration and conditions, the observed yields were in the range of 40 50%, with a higher yield with increasing flow rate. Thus, a 2-fold increase of crystal yield can be obtained by providing a flow to the crystallizing solution in the range of 6 16 mm/min. It has been reported that at higher flow rates, the effect of impurities in the solution appears to affect the process more than at lower flow velocities.19 Also, the relative improvement of the effect of flow on mass transfer would be lesser when we compare the increase from static to a low flow rate and from a low flow rate to higher flow rates. Combined with these effects, there would be a finite flow velocity where the crystal yield will be at its maximum and then decreases with increasing flow rate. Clearly, this is above the flow velocity that has been utilized in this study.

’ CONCLUSION We report a methodology for increasing nucleation and yield of insulin crystals using an oscillatory flow in capillary tubes. By utilizing the intermittent flow modes, we were able to segregate the effect of oscillatory flow on nucleation and crystal growth. This method has a great advantage over unidirectional flow due to the influence on nucleation behavior of the protein. Thus, a

ARTICLE

unique method to control the number of crystals that can be generated in protein crystallization has been put forward by this study. The influence is clearly not just on the number of the crystals as the crystal yield has increased from 24% under stationary conditions to 50% under the highest oscillatory flow rate of 12 mm/min used in this study. Thus, the oscillatory flow can be utilized to produce a large number of insulin crystals with a size less than 50 μm while maintaining a higher yield of product. This potential has been demonstrated through our study as a number of crystals with about 50 μm size are observed in the microscope pictures of crystals growing under oscillatory flow conditions for 48 h. Since the flow rate and the frequency of the intermittent oscillatory flow can be adjusted independently, it offers a wider opportunity to analyze and control the events of crystallization. This method is also applicable to a wider spectrum of shear- and temperature-sensitive proteins that cannot be crystallized under continuous flow modes due to process difficulties and protein denaturation. The flow modes in capillary crystallization can potentially be optimized to enhance nucleation, increase crystal yields, and control crystal size and can potentially be applied for downstream separation and formulation of high value products.

’ AUTHOR INFORMATION Corresponding Author

*Tel: +44 (0)20 7594 0784. Fax: +44 (0)20 7594 5700. E-mail: [email protected].

’ REFERENCES (1) Shan, N; Zaworotko, M. J. The role of cocrystals in pharmaceutical science. Drug Discovery Today 2008, 13, 440–446. (2) Vishweshwar, P; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J. Pharmaceutical co-crystal. J. Pharm. Sci. 2006, 95, 499–516. (3) McPherson, A. Introduction to protein crystallization. Macromol. Cryst. 2004, 34, 254–265. (4) Shenoy, B; Wang, Y; Shan, W; Margolin, A. Stability of crystalline proteins. Biotechnol. Bioeng. 2001, 73, 358–369. (5) Margolin, A; Navia, M. Protein crystals as novel catalytic materials. Angew. Chem. (Int. Ed. Engl.) 2001, 40, 2205–2222. (6) Rosenberger, F. Inorganic and protein crystal-growth - Similarities and differences. J. Cryst. Growth 1986, 76, 618–636. (7) Segelke, B. Efficiency analysis of sampling protocols used in protein crystallization screening. J. Cryst. Growth 2001, 232, 553–562. (8) Hursthouse, M; Huth, L; Threlfall, T. Why do organic compounds crystallise well or badly or ever so slowly? Why is crystallisation nevertheless such a good purification technique? Org. Process Res. Dev. 2009, 13, 1231–1240. (9) DeLucas, L. J. Protein crystallization - is it rocket science? Drug Discovery Today. 2001, 6, 734–744. (10) Derewenda, Z; Vekilov, P. Entropy and surface engineering in protein crystallization. Acta Crystallogr. Sect. D 2006, 62, 116–124. (11) Basu, S; Govardhan, C; Jung, C; Margolin, A. Protein crystals for the delivery of biopharmaceuticals. Exp. Opin. Biol. Ther. 2004, 4, 301–317. (12) Wang, M; Yin, X; Vekilov, P; Peng, R; Ming, N. Intrinsic instability of the concentration field in diffusion-limited growth and its effect on crystallization. Phys. Rev. E 1999, 60, 1901–1905. (13) Schruben, D; Stephanus, J; Gonzalez, M. Crystal growth mechanisms, transport enhanced crystallization. Cryst. Growth Des. 2009, 9, 2794–2800. (14) Murai, R; Yoshikawa, H; Kawahara, H; Maki, S; Sugiyama, S; Kitatani, T. Effect of solution flow produced by rotary shaker on protein crystallization. J. Cryst. Growth 2008, 310, 2168–2172. (15) Hou, D; Chang, H. AC Field enhanced protein crystallization. Appl. Phys. Lett. 2008, 92, 223902–223904. 4358

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359

Crystal Growth & Design

ARTICLE

(16) Ramachandran, N; Leslie, F. Using magnetic fields to control convection during protein crystallization-analysis and validation studies. J. Cryst. Growth 2005, 274, 297–306. (17) Lorber, B. The crystallization of biological macromolecules under microgravity: a way to more accurate three-dimensional structures? Biochim. Biophys. Acta, Proteins Proteomics 2002, 1599, 1–8. (18) Nanev, C. N.; Penkova, A; Chayen, N. Effects of buoyancydriven convection on nucleation and growth of protein crystals. Ann. N. Y. Acad. Sci. 2004, 1027, 1–9. (19) Roberts, M. M.; Heng, J. Y. Y.; Williams, D. R. Protein crystallization by forced flow through glass capillaries: enhanced lysozyme crystal growth. Cryst. Growth Des. 2010, 10, 1074–1083. (20) Kurzweg, U. H. Enhanced diffusional separation in liquids by sinusoidal oscillations. Sep. Sci. Technol. 1988, 23, 105–117. (21) Murray, B. T.; Coriell, S. R.; Chernov, A. A.; McFadden, G. B. The effect of oscillatory shear flow on step bunching. J. Cryst. Growth 2000, 218, 434–446. (22) Ristic, R. I. Oscillatory mixing for crystallization of high crystal perfection pharmaceuticals. Chem. Eng. Res. Des. 2007, 85, 937–944. (23) Rujano, J. R.; Crane, R. A.; Rahman, M. M.; Moreno, W. Numerical analysis of stabilization techniques for oscillatory convective flow in Czochralski crystal growth. J. Cryst. Growth 2002, 245, 149–162. (24) Abel, K. Influence of oscillatory flows on protein ultrafiltration. J. Membr. Sci. 1997, 133, 39–55. (25) Adachi, H; Takano, K; Niino, A; Matsumura, H; Kinoshita, T; Warizaya, M. Solution stirring initiates nucleation and improves the quality of adenosine deaminase crystals. Acta Crystallogr. Sect. D 2005, 61, 759–762. (26) Yaoi, M; Adachi, M; Takano, K; Matsumura, H; Inoue, T; Mori, Y. Effect of stirring method on protein crystallization. Jpn. J. Appl. Phys. 2004, 43, L1318–L1319. (27) Bergeron, L; Filobelo, L; Galkin, O; Vekilov, P. Thermodynamics of the hydrophobicity in crystallization of insulin. Biophys. J. 2003, 85, 3935–3942. (28) Penkova, A; Pan, W; Hodjaolgu, F; Vekilov, P G. Nucleation of protein crystals under the influence of solution shear flow. Ann. N.Y. Acad. Sci. 2006, 1077, 214–231. (29) Haw, M D; Poon, W C K; Pusey, P N. Direct observation of oscillatory-shear-induced order in colloidal suspensions. Phys. Rev. E 1998, 57 (6), 6859–6864. (30) Slawecki, T M; Glinka, C J; Hammouda, B. Shear-induced micellar crystal structures in an aqueous triblock copolymer solution. Phys. Rev. E 1998, 58 (4), R4084–R4087. (31) Kadima, W; McPherson, A; Dunn, M F; Jurank, F. Precrystallization aggregation of insulin by dynamic light scattering and comparison with canavalin. J. Cryst. Growth 1991, 110, 188–194. (32) Gliko, O; Neumaier, N; Fischer, M; Haase, I; Bacher, A; Weinkauf, S; Vekilov, P G. Dense liquid droplets as a step source for the crystallization of lumazine synthase. J. Cryst. Growth 2005, 275, e1409–e1416. (33) Vermant, J; Solomon, M. J. Flow-induced structure in colloidal suspensions. J. Phys.: Condens. Matter 2005, 17 (4), R187–R216.

4359

dx.doi.org/10.1021/cg200158z |Cryst. Growth Des. 2011, 11, 4353–4359